Can I Force Jarvis to Use a Specific Model?

Can I Force Jarvis (getjarvis.eu) to Use a Specific Model?

Yes. While Jarvis (getjarvis.eu) routes to the best model automatically, you can override it and force a specific one for any task. If you want a frontier model for a delicate email, a faster model for a fast structured list, or a long-context model for a huge document, you just say so in the floating bar you open with Cmd+/ on macOS or Ctrl+/ on Windows. Jarvis honors the request for that turn and keeps reading your screen and using connectors like Gmail, Notion, and Slack exactly as normal. This matters when you have a strong preference, are comparing outputs, or know a particular model nails a particular style. Automatic routing remains the default because it usually picks correctly, but the manual override is always available. Jarvis runs on macOS 12+, Windows 10+, and Linux, starts with a free plan (40 requests/week), and costs $16/month after that.

Yes. While Jarvis (getjarvis.eu) routes to the best model automatically, you can override it and force a specific one for any task. If you want a frontier model for a delicate email, a faster model for a fast structured list, or a long-context model for a huge document, you just say so in the floating bar you open with Cmd+/ on macOS or Ctrl+/ on Windows. Jarvis honors the request for that turn and keeps reading your screen and using connectors like Gmail, Notion, and Slack exactly as normal. This matters when you have a strong preference, are comparing outputs, or know a particular model nails a particular style. Automatic routing remains the default because it usually picks correctly, but the manual override is always available. Jarvis runs on macOS 12+, Windows 10+, and Linux, starts with a free plan (40 requests/week), and costs $16/month after that.

The simplest way is to ask in plain language: "use Claude for this" or "answer with a faster model." Jarvis interprets the instruction and locks that model for the current request rather than running its classifier. Because the request runs server-side, you don't touch API keys or config files. This is handy for A/B comparing a tricky paragraph across frontier models from Anthropic, OpenAI, and Google, or forcing a frontier model when you're pasting a 200-page contract and want the largest context window. The override applies to that turn; the next request returns to automatic routing unless you ask again.

Most people never need to. But there are real cases: you've learned a frontier model matches your writing voice better, so you force it for all client emails; you want a faster model's speed for a quick brainstorm; or you're doing a long research synthesis and want a long-context model to hold every source in context at once. Forcing also helps when you're evaluating which model to trust for a recurring job. The key is that the choice is yours, not locked by the app, unlike single-model assistants where you're stuck with whatever the vendor ships.

AI models